juicebox notebook module

Jupyter Notebook

JupyterLab

=======

juicebox-notebook is an module for Jupyter and Colab notesbooks which exposes a python API that enables creation and interaction with a juicebox.js instance in a cell.

Examples

Example notebooks are available in the github repository, and can be run from the Binder and Colab links above. To download examples without cloning the repository use this link. Notebooks are available in the "examples" directory.

Installation

Requirements:

• python >= 3.6.4
• jupyter >= 4.2.0

pip install juicebox-notebook

Initialization

After installing import and intialize igv_notebook as follows.

import juicebox_notebook
juicebox_notebook.init()

For a Jupyter notebook this should be done once per notebook. Colab notebooks display output in a sandboxed iFrame for each cell, so these steps must be repeated for each cell in which juicebox-notebook is used.

Browser creation

The juicebox.Browser initializer takes a configuration dictionary which is converted to JSON and passed to the juicebox.js init function. See the juicebox.js README

Example:

browser = juicebox_notebook.Browser(
    {
        "name": "GM12878",
        "url": "https://www.encodeproject.org/files/ENCFF179HVU/@@download/ENCFF179HVU.hic",
        "tracks": [
            {
                "url": "https://www.encodeproject.org/files/ENCFF000ARJ/@@download/ENCFF000ARJ.bigWig",
                "name": "CTCF",
                "color": "rgb(22, 129, 198)"
            }
        ]
    }

)

The juicebox_notebook.Browser initializer takes a configuration object which is converted to JSON and passed to the juicebox.js createBrowser function. The configuration object is described in the juicebox.js documentation. To open an empty "browser" to dynamically load maps pass an empty dictionary

Loading Maps and Tracks

Typically maps and tracks are loaded in the initial juicebox.Browser creation. However its also possible to load them post creation using the b.load_map(config and b.load_track functions.

Maps

To load a map into an existing browser pass a hic file configuration object to the load_map function

import juicebox_notebook
juicebox_notebook.init()
b = juicebox_notebook.Browser({})
b.load_map(
    {
         "url": "https://hicfiles.s3.amazonaws.com/hiseq/gm12878/in-situ/primary.hic"
    }
)

Tracks

To load a track pass a track configuration object to load_track. Track configuration objects are described in the juicebox.js documentation. The configuration object will be converted to JSON and passed to the juicebox.js browser instance.

import juicebox_notebook
juicebox_notebook.init()
b = juicebox_notebook.Browser({
        "name": "GM12878",
        "url": "https://www.encodeproject.org/files/ENCFF179HVU/@@download/ENCFF179HVU.hic"
    }
)

b.load_track({
    "url": "https://www.encodeproject.org/files/ENCFF000ARJ/@@download/ENCFF000ARJ.bigWig",
    "name": "CTCF",
    "color": "#ff8802"
  }
)

URLS and paths

Configuration objects for juicebox.js maps (.hic files) and tracks have properties to specify URLs to files for data and indexes. These properties are supported in juicbox-notebook, however juicebox-notebook also provides equivalent path properties for specfiying paths to local files when used with Jupyter Notebook or Colab.
(Note: The path properties cannot be used with JupyterLab, however local files can be loaded by URL). The path properties are useful for

1. Loading data in a Colab notebook from the local Colab file system or a mounted Google Drive
2. Loading data in Jupyter Notebook from the local file system that is outside the Jupyter file tree.

For Jupyter servers (Notebook and Lab) local files can be also be loaded via the url property if the file is in Jupyter directorytree. This will usually yield better performance than using path properties. URLs that begin with a "/" are relative to the Jupyter server startup directory, that is the directory from where you started Jupyter Notebook or JupyterLab. URL paths without a leading slash are assumed to be relative to the notebook directory.
See below for examples. You can also use the "download url" for the file, obtainable through the JupyterLab UI, as the URL for juicebox.

URL and Path properties | juicebox.js url property | juicebox-notebook path property | notes | --------- | ----------- | | url | path | | indexURL | indexPath | Used in some track configurations. See igv.js

Examples:

Local files using paths - Jupyter and Colab. Absolute file paths, potentially outside the Jupyter file tree. Note the use of path instead of url.

import juicebox_notebook

juicebox_notebook.init()
b = juicebox_notebook.Browser(
    {
        "name": "GM12878",
        "path": "/TestData/juicebox/HCT-116_Untreated.hic",
        "tracks": [
            {
                "path": "/TestData/juicebox/CTCF_Untreated.bw",
                "type": "wig",
                "name": "CTCF",
                "color": "rgb(22, 129, 198)"
            }
        ]
    }
)

Local files using absolute urls - Jupyter. The URL is relative to the startup directory of the Jupyter server

import juicebox_notebook

juicebox_notebook.init()
b = juicebox_notebook.Browser(
    {
        "name": "GM12878",
        "path": "/TestData/juicebox/HCT-116_Untreated.hic",
        "tracks": [
            {
                "path": "/TestData/juicebox/CTCF_Untreated.bw",
                "type": "wig",
                "name": "CTCF",
                "color": "rgb(22, 129, 198)"
            }
        ]
    }
)

Local files using relative urls - Jupyter. URL relative to the directory of the notebook

import juicebox_notebook

juicebox_notebook.init()
b = juicebox_notebook.Browser(
    {
        "name": "GM12878",
        "path": "TestData/juicebox/HCT-116_Untreated.hic",
        "tracks": [
            {
                "path": "TestData/juicebox/CTCF_Untreated.bw",
                "type": "wig",
                "name": "CTCF",
                "color": "rgb(22, 129, 198)"
            }
        ]
    }
)

Development notes

Creating a conda environment:

conda create -n juicebox_notebook python=3.9.1
conda activate juicebox_notebook
conda install pip
conda install jupyter

Build and install from source:

python setup.py build  
pip install -e .